Catch me if you can!
This story of Mu Sigma’s engagement with one of Australia’s largest financial services brands is inspirational and encompasses lessons worth learning from.
Mu Sigma team started its engagement with the client solving smaller problems related to their individual business units.
An organizational restructuring in the client’s system, and the setup of the Chief Data Office catapulted the prominence of what we were doing. After a grueling RFP process, we were chosen as their Advanced Analytics Partner.
The Mu Sigma team along with the Chief Data Office of the client worked towards improving the efficiency and effectiveness of their existing Fraud Management System.
Problem and scope of the engagement…
The existing fraud management system was managed by a third party vendor. The client was not satisfied with the performance and trusted us to plug the gaps and improve the overall system.
There were some major challenges we faced!
The volume of claims being handled by the system was massive. A scalable solution was required in order to streamline the process.
We connected with multiple stakeholders within the client ecosystem to understand their functioning and key pain points. We then identified opportunities across the system, and charted out an analytical roadmap to improve the system.
The existing mechanism of fraud identification was done by a third party vendor using a proprietary, black-box algorithm. Lack of transparency in the process increased the dependency on the vendor, in addition to the costs incurred for fraud identification. We proposed developing our own fraud identification system that would reside in-house with the added flexibility of incorporating new data sources.
While implementing our model we were to capture characteristics that were to be indicative of suspicious behavior. This was a challenge, compounded by the fact that such behavior evolve with time.
We employed first principles thinking and arrived at different hypotheses regarding characteristics of fraudulent claims. We also incorporated the decision tree provided by the stakeholders to ensure all human experiences are taken into account during the process.
This was followed by a comprehensive list of variables suggestive of fraudulent transactions.
We also conducted bi-weekly catch up with the stakeholders which maintained the continuous feedback loop.
Finally, as a result of this systematic and inclusive approach, we were able to come up with a robust solution design.
Our approach and success
The systematic approach of identifying fraudulent transaction led to the development of an improved and effective system
Following the success at fraud identification on historical data, the model was deployed for test and learn on real time data. We have realized actual savings of around $200,000 in just 2 months for a subset of claims. The system is being further improved by incorporating new data sources and techniques.
Once the potential cost savings are realized, the system is then slated to be deployed full-scale, replacing the existing system.
Learning from the ecosystem
We had focused discussions with other accounts within Mu Sigma to understand where the key pain points lie. A preemptive muPDNA was formed based on our past experience with insurers and first principal thinking. This allowed us to have richer and more engaging discussions with the stakeholders.
We also suggested some of the techniques used in credit risk modelling to the client. Our approach of going beyond the stated problem and being creative while identifying solutions was much appreciated by the client.
Client’s appreciation – Excerpts
The client is very appreciative of our approach towards problem solving. They were impressed by the fact that we look at problems from an end to end point of view unlike other vendors.
The fraud management project was a complex project to work on due to a lot of dynamic factors. They were impressed by the results we achieved. Our success was also suggestive of the good partnership we had with the client’s CDO team.
Trends that are catching up – in the Banking and Insurance sector
Organizations are using technology and big data for commercial success. Social media data, telematics, biometrics etc. are being employed to understand the customers better.
The market is highly competitive and organizations are focusing on creating convenient and easy ways for customers to do business with them. They want to be the one stop solution for all banking and insurance services for the customers – convergence will be the way forward.
The old agent based traditional sales model is now diminishing. Customers are getting more informed due to the digital revolution. Hence there is a rise of online players including insurers, marketplaces and aggregators.
Like what you are reading? Read some more interesting blogs and case studies on Our Musings section.